Latest posts tagged with #MedAI on Bluesky
Big tech is moving into your medical records. Microsoft’s Copilot Health will soon merge Apple Watch data with clinical histories to "connect the dots" in seconds. It’s an efficiency win, but the privacy stakes are high: HIPAA doesn't apply here, and AI bias hasn't been solved yet.
#MedSky #MedAI
However, prospective tests showed AI needs constant tuning. A shift in hardware (new scanners) doubled the recall rate, requiring a threshold recalibration.
AI isn't set and forget; it's a tool that needs active, localized maintenance to keep screening fair and safe.
#MedSky #MedAI
What's impressive is the scale. By simulating AI as a second reader, the team saw a 32% drop in radiologist time and a 17% jump in cancer detection. It's especially strong for first-time screens where there's no prior image to compare.
#MedSky #MedAI
A new massive study of breast cancer screening AI (115,000+ UK cases) is out. Google's mammography AI system didn't just match radiologists, it beat the first reader on sensitivity and caught 25% of interval cancers humans missed.
#MedSky #MedAI
Integrating temporally localized notes into sepsis models boosted predictive performance to an AUC of 0.84, significantly outperforming traditional models using only structured numerical data (AUC 0.77).
#MedSky #MedAI
LLMs can reliably identify and time patient outcomes in unstructured clinician notes. Meta-Llama-3.1-8B hit the best balance of precision and coverage, while the biomedical-focused BioMistral 7B tended toward over-prediction. Contextual reasoning beats exact term matching.
#MedSky #MedAI
New Review 💬 Generative Artificial Intelligence Methodology Reporting in Otolaryngology
🔗 onlinelibrary.wiley.com/doi/10.1002/...
#BiomedicalInformatics #MedSky #MedAI #AISky #AI #ML #GenerativeAI #ChatGPT #GoogleGemini #Claude #ArtificialIntelligence #GenAI #LLMs
Bridge2AI March Events: Tuesday, March 17th (12:00 PM PT/3:00 PM ET): Bridge2AI Discussion Forum on Emerging ELSI Issues with Dr. Nicholas Evans: Conjoint Analysis of Perspectives on Ethical Tradeoffs in Data Generation Project Thursday, March 19th (12:00 PM PT/3:00 PM ET): Bridge2AI TRM Novel AI Technology Module Lecture with Dr. William Speier: Multimodal Deep Learning Models for Thyroid Cancer Risk Stratification
Join us for our upcoming seminars in #ArtificialIntelligence 🚀
1️⃣Conjoint Analysis of Perspectives on Ethical Tradeoffs in Data Generation Project
2️⃣Multimodal Deep Learning Models for Thyroid Cancer Risk Stratification
🔗 bridge2ai.org/events/
#BiomedicalInformatics #MedSky #MedAI #AISky #AIML
Healthcare is shifting from "shadow AI" to secure, managed tools, with personal app usage dropping from 82% to 32%. Yet data risks remain high, regulated data accounts for 89% of policy violations.
#MedSky #MedAI
Silent brain infarctions are easy to miss on serial MRI, but they're a huge warning sign. A new model with an 0.89 AUC found that "silent" lesions are linked to a 3.8-fold increased risk of a future stroke and could be a better way to manage risk.
#MedSky #MedAI
Merlin is a new 3D AI model for abdominal CTs that’s actually holding up across different hospitals. Usually, these models tank when they see data from a new site, but Merlin averaged a 19.7% jump in accuracy over the next best system across three external centers.
#MedSky #MedAI
Catalyzing Health AI by Fixing Payment Systems
"In this article, we examine the reimbursement landscape for health #AI, focusing first on tools that fit existing regulatory pathways, outlining payment barriers and proposing policy reforms."
🔗 pubmed.ncbi.nlm.nih.gov/41695240/
#MedSky #MedAI
New Publication: Quantum Machine Learning and Data Re-Uploading: Evaluation on Benchmark and Laboratory Medicine Data Sets 🚀
🔗https://pubmed.ncbi.nlm.nih.gov/41728802/
#ArtificialIntelligence #BiomedicalInformatics #MedSky #AISky #MedAI #QML #AIML #QuantumMachineLearning
Medical AI agents show modest accuracy gains but use 10x more tokens and double the response time. Hallucinations still do occur, which means these systems are probably not yet ready for prime time yet.
Keep in mind GPT-4.1, Qwen-3 are older models, but Llama 4 is the current model.
#MedSky #MedAI
LLM simplify radiology reports to aid patient understanding.
A review of 38 studies shows patients perceived LLM-rewritten reports as significantly more understandable than radiologist reports (Likert scores 2.16 to 4.04). Doctors found them accurate despite a small error risk.
#MedSky #MedAI
#YOLO isn't just a motto, it's a model! 🚀 Explore how You Only Look Once was used to develop a #deeplearning object detection system for identifying vocal fold polyps in stroboscopic video frames: onlinelibrary.wiley.com/doi/10.1002/...
#ArtificialIntelligence #BiomedicalInformatics #MedSky #MedAI
Want to learn more about the Bridge2AI program and our data generation projects addressing grand challenges? Visit: bridge2ai.org 🚀
#ArtificialIntelligence #BiomedicalInformatics #MedSky #AISky #MedAI #AI #ML
The AIREADI dataset uses data from participants with and without T2DM, collected via methods like surveys and clinical measurements at 3 sites. The data is hosted in Fairhub and accessible under controlled access.
🔗 aireadi.org/dataset
#ArtificialIntelligence #BiomedicalInformatics #MedSky #MedAI
The Voice dataset uses an app to collect acoustic tasks, surveys, and questionnaires from participants at 5 North American sites. The data is hosted in PhysioNet with registered and controlled access to preserve voice data.
🔗 b2ai-voice.org/bridge2ai-vo...
#BiomedicalInformatics #MedSky #MedAI
The CM4AI dataset uses mapping techniques to generate a library of large-scale cell maps. The data is stored in RO-Crate packaging and hosted in UVA Dataverse.
🔗 cm4ai.org/data-releases/
#ArtificialIntelligence #BiomedicalInformatics #MedSky #AISky #MedAI #AI #ML
The CHoRUS dataset uses data collected from patient admissions to ICUs across 14 hospitals in the U.S. The data is hosted in MGB Azure with access managed through registration. 🔗 chorus4ai.org/dataset/
#ArtificialIntelligence #BiomedicalInformatics #MedSky #AISky #MedAI #AI #ML
AI deployment in primary care lacks standards for clinical outcomes. PRIMARY-AI aims to fill this gap by establishing a framework to measure people-centered results and equity.
#MedSky #MedAI
Knowing what you have is less important than knowing what to do.
Even when LLMs helped identifying conditions, they didn't improve decisions on seeking care (e.g., "go to ER" vs. "wait"). AI advice must drive the correct actions, not just provide information.
#MedSky #MedAI
Why do powerful models fail in practice? Poor prompting.
Users often treat medical chatbots like search engines, asking vague questions that yield generic advice.
Without structured guidance, the "intelligence" of the model is lost in translation.
#MedSky #MedAI
LLMs ace medical exams, but that doesn't make them good doctors.
In a Nature Medicine study of 1,298 people, users paired with AI identified conditions correctly <35% of the time, far worse than the AI alone (95%). Showing that access to knowledge isn't the same as effective care.
#MedSky #MedAI
New in Nature Health “A foundation model for breast and lung cancer screening using non-contrast computed tomography.”
Transform from single-test single-cancer to multi-cancer screening.
www.nature.com/articles/s44...
#MedAI
New publication in Wiley Laryngoscope: Exploring Real-Time Tracking of Vocal Fold Polyps in Video-Stroboscopy Using #DeepLearning 💡 onlinelibrary.wiley.com/doi/10.1002/...
#AI #ArtificialIntelligence #BiomedicalInformatics #MedSky #MedAI #MLSky #VoiceAI
🧪 A new study finds that general cardiologists using the AMIE AI system reduced clinical errors from 24.3% to 13.1%.
Sub-specialists preferred assisted care 46.7% of the time.
The model saved time in 50.5% of cases and helped bridge expertise gaps.
#MedSky #MedAI
